From: Maxim Levitsky <mlevitsk@redhat.com>
Date: Thu, 29 Jul 2021 18:37:38 +0300
Subject: KVM: nSVM: always intercept VMLOAD/VMSAVE when nested
Patch-mainline: No, still under Embargo
References: bsc#1189400, CVE-2021-3656

If L1 disables VMLOAD/VMSAVE intercepts, and doesn't enable
Virtual VMLOAD/VMSAVE (currently not supported for the nested hypervisor),
then VMLOAD/VMSAVE must operate on the L1 physical memory, which is only
possible by making L0 intercept these instructions.

Failure to do so allowed the nested guest to run VMLOAD/VMSAVE unintercepted,
and thus read/write portions of the host physical memory.

This fixes CVE-2021-3656, which was discovered by Maxim Levitsky and
Paolo Bonzini.

Fixes: 89c8a4984fc9 ("KVM: SVM: Enable Virtual VMLOAD VMSAVE feature")
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Acked-by: Joerg Roedel <jroedel@suse.de>
---
 arch/x86/kvm/svm.c |    3 +++
 1 file changed, 3 insertions(+)

--- a/arch/x86/kvm/svm.c
+++ b/arch/x86/kvm/svm.c
@@ -484,6 +484,9 @@ static void recalc_intercepts(struct vcp
 	c->intercept_dr = h->intercept_dr | g->intercept_dr;
 	c->intercept_exceptions = h->intercept_exceptions | g->intercept_exceptions;
 	c->intercept = h->intercept | g->intercept;
+
+	c->intercept |= (1ULL << INTERCEPT_VMLOAD);
+	c->intercept |= (1ULL << INTERCEPT_VMSAVE);
 }
 
 static inline struct vmcb *get_host_vmcb(struct vcpu_svm *svm)